This repository is an implementation of travel based recommendation system. This is an attempt to recommend places by including following features:

• Past and Current weather data
  
• Twitter popularity
  
• TripAdvisor reviews
  

https://youtu.be/RrdsDVKTWQU

• Scrapped TripAdvisor's reviews for selected places from 4 states: Colorado, California, Arizona, New York
  
• Collected tweets for the same places using Twitter REST API
• Collected weather data on a monthly basis averaged over past 12 years from WorldWeatherOnline API
• Collected one recent image per location by scraping Instagram
• Generated latitude and longitude for these places using OpenCage API

• Created various collections for each kind and stored the above mentioned dataset in MongoDB Atlas cluster hosted on GCP
• Used VADER for sentiment analysis of TripAdvisor's reviews and generated an overall score for a place by giving importance more recent reviews and reviews having more upvotes.
• Our algorithm gives more weightage to tweets over reviews from TripAdvisor.